Introduction to the IPUMS Translation Tables

Go Back to IPUMS Documentation Index

The integration of disparate variable names and coding schemes is one of the most important goals of the IPUMS project. "Translation tables" are the main tool that we use to integrate various coding schemes over time. The basic function of a translation table is to show how codes from the original Public Use Microdata Samples align to codes in the IPUMS files that we make available on this website.

Translation tables have a consistent format. The first two columns of each translation table show the "integrated" IPUMS code and the IPUMS label.  Subsequent columns show how values in the original source data correspond to the IPUMS codes.

Each of the columns describing original source data includes cells identifying the sample (e.g. us1940a), record type ("H" or "P"), and variable location in the original data (e.g., column 80).  The codes in each column on the right-hand side of the translation table are values from the original source data. Each value in the original source data is aligned in the same row as its integrated IPUMS code in the leftmost column. 

If samples from from multiple years share a record layout and coding scheme (as is often the case in the American Community Survey samples), those samples will share a column in the translation table.  The same is true when multiple samples exist for a given year (for instance, the 2000 1% and 5% samples): the samples will share a column if they have the same codes and layout; they will have separate columns if they have different codes or layouts.

The list at the bottom of this page shows how samples are identified in the translation tables. 

Other details:

• Some variables do not require any recoding; these are identified with a value of "1" in the table row lablelled "norecode". 
• An asterisk (*) in any cell indicates that for that year cases with the IPUMS value for that row were created by special programming rather than through simple recoding. Sometimes entire variables need to be created via complex programming (rather than simple recodes); in those cases the original column location is listed as "0".
• A "B" or "BB" or any combination of "B" indicates values in the input data that were left blank.

Below is a modified example of the translation table for the IPUMS variable VETSTAT that demonstrates some of the features of the translation tables.

code	label	us1940a	us1950a	us1960a	us1970b	us1980a	us1990a	us2000a	us2000d us2001a us2002a	us2005a us2006a
rectype		P	P	P	P	P	P	P	P	P
columns		164=165	0	80	80	72	72	138	53	60
norecode			1
0	N/A			2	2	*	0
0	Under age 18	18
0	Under age 18	21
0	Under age 18	22
0	Under age 18	28
0	Wife of veteran	23
0	Under age 17							0	B	B
0		BB
1	No Service	19		1	1	1	4	4	4	5
1	Now on active duty						1	1	1
1	Training for Reserves or National Guard only									4
2	Yes	29		0	0	0
2	Yes, on active duty in past but not now						2	2	2	2
2	Yes, service in reserves or national guard only						3	3	3	3
9	Unknown	88
9	Not ascertained	89

Original "input" or "source data" samples listed in the translation tables